Dairy is big business in Canada, with some 1.4 million cows, most of them in Québec and Ontario, responsible for $6.17 billion in net revenues each year. Dairy in Canada is also unique in that milk production is subject to a quota system and supported by more than 13,000 farms, mostly family-run. Optimizing productivity, maximizing resources and limiting expenses are directly associated with farm profitability. Innovations such as genetic selection, increased management efforts and control over production variables have helped maximize profitability during the past several years. Production data such as milk yield, milk components, animal characteristics and management conditions are available for the past 40 years but have never been integrated with genomic data.
Abdoulaye Baniré Diallo of the Université du Québec à Montréal and Marc-André Sirard of Université Laval are developing tools that will perform this integration, driving the development of new management practices and allowing a precise lifetime productivity estimate for individual cows. Their data mining and machine learning toolkits will deliver predictive models of dairy production that will influence management practices and optimize dairy farm profitability. The project takes advantage of the advances in genomics that allow an increasingly detailed genetic profile to be established for each individual dairy cow, at relatively low cost, optimizing decisions about feeding, reproduction, therapeutic treatments and caregiving, and will be applicable to farms across Canada. It will also be a stepping stone to including new selection traits such as response to various housing conditions, heifer growth and adaptation to robotized milking systems, to breeding programs. This research is performed with a multisectorial and transdisciplinary research team composed of scientists from Université du Québec à Montréal, Université Laval and McGill University, in partnership with Valacta and My Intelligent Machines (Mims).